Wireless Positioning Based On Kalman Filter In NLOS Environment

Wireless positioning, performed either in network or in mobile station, is to determine the geographic location of mobile station through the estimated time and angle measurements with the help of location algorithms. The researchers'passion on this field was stimulated by E-911 location regulation submitted by FCC in 1996 and promoted further by the inexhaustible commercial potential. More and more scholars devote themselves to wireless positioning filed, and large numbers of research results and patents have appeared.Two main modules are required in an entire location process:1) estimation of the location parameters and 2) operation of the location algorithms. The time or angle information, which is indispensable to location, is abstracted from the received signal through some parameter estimation algorithms and then input to the location module to yield the location estimation. The primary difficulties lie in the resistance against the estimating errors caused by the severe wireless channel, and the solution to the nonlinear equations which exhibit the geometric relation between the mobile station and the participant base stations.We have stated some classic location algorithms such as Least Square method, Chan's algorithm, Taylor Series Expansion method and Wylie algorithm, especially given a detailed introduction of Kalman filter-based location method. These methods all have their own characteristics, the location performance of Chan's algorithm is significantly degraded in non-line-of-sight (NLOS) propagation environment. Wylie algorithm, reconstruction of range measurements and the tracking of Kalman filter method can mitigate the NLOS error in time measurements effectively. Especially taking advantage of the tracking ability of Kalman filter, this method can provide high accuracy and robustness to the variation of the measured parameters. Besides, we introduce NLOS model and its property, and NLOS error identification and mitigation.We made an intensive study aiming at location technique in NLOS environment and proposed two NLOS error mitigation schemes. The main contributions of the paper are summarized as follows:1. We proposed a robust NLOS error mitigation method in mobile positioning. The method mitigates NLOS error in raw measurements by using the tracking characteristic of Kalman filter and reconstructs true measurements by using range reconstruction algorithm.2. We proposed a weighted least square location algorithm based on Kalman filter in NLOS environment. First, range measurements are identified according to their statistical characteristics of standard deviation of measurement noise and a weight factor is assigned to each corresponding measurement. Then two-step Kalman filter is used to track the raw measurements and new measurements are reconstructed using the weight factor. Finally, a weighted least square method is proposed to give the location estimation. Through simulation analysis and performance comparison, our proposed two location algorithms can effectively NLOS error in TOA measurements and have higher precision.In conclusion, we have been trying to summarize the methodology of the research on wireless positioning, trying to tamp the base so as to sparkle the new idea.